Utah Cost of Service and Rate Design Task Force

1. UtahCost of Service and Rate DesignTask Force Committee of Consumer Services Tony Yankel August 25, 2005

2. OVERVIEW • System vs. Jurisdictional cost allocation • Retail load pattern • Generation pattern • Long-term Firm Wholesale pattern • Short-term Firm Wholesale pattern • Economy Wholesale pattern • Cost allocation implications • Customer Charge

3. System vs. Jurisdictional Costs • Costs come from system load, not isolated jurisdictional loads. • Costs come from the way system costs are allocated to the jurisdiction. • The appropriate starting place for defining costs is to look at the system load and allocation factors.

4. Retail Load Pattern(MW x 1,000)

5. Retail Load Pattern(Percentage of Peak)

6. Retail usage patternsObservations • Load profile is relatively flat. • The non-6x16 usage (Sundays and nights) is lower, but not significantly lower than 6x16. • The 6x16 pattern is relatively flat for the monthly average as well as the peak day.

7. Average July 2003 6x16 Load

8. Average July 2003 6x16 Load Observations • The average July 6x16 Retail load shows and expected pattern of lower load at the ends of the period and high in the middle. • The peak 6x16 Retail load is 2,200 Mw’s or 1/3rd greater than the load at 7:00 and 900 Mw’s greater than at 22:00. • How are resources combined to meet Retail load?

9. Generation Pattern(MW x 1,000)

10. Generation Pattern(Percentage)

11. Generation PatternObservations • Generation is very flat, even with the use of Peaking resources. • Maximum generation did not occur at the time of the monthly coincident peak. • Approximately 5% of the hourly generation was greater than during the time of the monthly coincident peak.

12. Retail Load less Generation(MW x 1,000)(Red values are negative)

13. Retail Load less GenerationObservations • For Retail load, the Company is generally an importer of power during HLH and and exporter during LLH. • Daytime on Sundays is like any other HLH—Retail greater than Generation. • During HLH, retail load can be 20-30% greater than generation. • What are wholesale transactions like?

14. LTF Purchases (MW)

15. LTF Sales (MW)(Red values are leaving the system)

16. Net LTF (MW)(Red values areleaving the system)

17. Net LTF 6x16

18. Net LTFObservations • PacifiCorp was a net exporter of LTF during every hour of July 2003. • Net LTF transactions add another 10% (8-900 Mw) to the 6x16 System load. • On average, LTF is more “peaky” than retail load—requiring more HLH input than that required for Retail alone.

19. Retail plus LTF less Generation(Mw)

20. STF Purchase (MW x 1,000)

21. STF Purchases 6x16

22. STF Purchases 6x16Observations • STF Purchases are 2.5 times greater during HLH than LLH with a well defined line of demarcation. • STF Purchases are significant in size with the HLH purchases being 40% of the Retail load and 45% of system generation. • STF Purchases are 25% of Retail load during LLH and 22% of system generation. • PacifiCorp has more generation than Retail load during LLH, but it purchases 1,300 Mw’s of STF.

23. STF Sales (MW x 1,000)(Red values leaving the system)

24. STF Sales 6x16

25. STF Sales 6x16Observations • STF Sales are relatively flat at 1,700 Mw. • STF Sales increase by 700 Mw at the beginning of the HLH. • The increase in STF Sales at 7:00 a.m. is only about 1/3rd of the difference between Retail load at peak and at 7:00 a.m. (2,200 Mw).

26. STF Net (MW x 1,000) (Red values leave the system)

27. STF Net (MW)

28. STF NetObservations • Net STF is negative during LLH and somewhat shaped during HLH. • Net STF is significant and follows the shape of the Retail plus LTF requirements. However, it is not of sufficient magnitude to fully make up the difference between Retail plus LTF requirements and Generation.

29. OS Purchases (Mw)

30. OS Purchases (Mw) 6x16

31. OS Purchases • OS purchases are random as would be expected. • OS purchases are insignificant when compared to Retail load that is in the 6-8,000 Mw range.

32. OS Sales (Mw)(Red values leave the system)

33. OS Sales (Mw) 6x16

34. OS SalesObservations • OS Sales (although larger than OS Purchases) are insignificant when compared to Retail load that is in the 6-8,000 Mw range. • There is more OS sales during daylight hours than at night.

35. OS Net (MW x 1,000) 6x16 (Red values leave the system)

36. OS Net July 2003 (net sales)

37. OS NetObservations • Net OS activity is insignificant when compared to Retail load that is in the 6-8,000 Mw range. • The Net OS activity results in increased sales during the peak hours.

38. Sum of LF, SF, and OS Sales(Red means leaving the system)(Mw x 1,000)

39. Retail, Wholesale Salesand Generation 6x16 Retail plus Wholesale Sales Retail Generation

40. Retail, Wholesale, GenerationObservations • Wholesale sales run 40-50% that of Retail. • A large portion of system resources are required/used to supply non-Retail firm requirements. • Wholesale sales have a similar pattern to that of Retail with little or no “valley filling”.

41. Retail, Wholesale, GenerationObservations (Cont.) • Generation is extremely flat when compared with Retail or Wholesale sales. • Generation is insufficient to serve Retail during HLH and during all hours of the combined Retail, LF, SF, OS sales. • The System is operated to meet far more than the Retail requirements.

42. Allocation and Cost Causation • Allocation factors based upon Retail load may not reflect overall system operations. • Generation is run fairly flat, so these costs should be heavily spread on the basis of energy. • Purchase power is used to meet System requirements, not just Retail.

43. Inadequacy of CP Allocation • Assume Group A and B have the same demand at peak. • A CP method would equally allocate fixed costs to each group. • However, Wholesale revenue would only be generated by Group B as Group A would have no resources to offer for off-system sales. Group B Group A

44. Normalized Net WholesaleMarch 2005 Semi-Annual Report

45. Conclusions RegardingResource Costs • PacifiCorp’s generation is used flat-out with an energy allocation factor looking like it would be representative of costs. • Total Retail load is not flat, but it is for some customer groups. • The difference in Retail load and Generation is addressed in the Wholesale market. • If a group is allocated the same resources at peak as it uses during the rest of the time (flat load), it has no stake in Wholesale revenues.

46. Customer Charge • Yes, some Costs are allocated on the basis of the number of Customers or Weighted Customers. • However, the Customer Charge is a rate design issue, not a cost allocation issue. • There are a number of customer groups allocated demand costs, that have a zero demand charge. (Rate design vs. Cost Allocation)

47. Benefits of aLow Customer Charge • Conservation: More emphasis is placed upon the energy charge—the charge the customers have control over. • Customer understanding: Residential customers are not charged for entering a grocery store, why pay one to the utility? • Low bill for low usage: The proportion of the bill for low usage customers is far more related to the amount of usage.

48. Benefits of aHigh Customer Charge • Utility has a higher fixed income. • Utility revenue is less dependant upon temperature sensitive load or conservation measures. • From a rate analyst point of view, revenue collection would be more in line with cost allocation methods.

49. Price Signals Associated withthe Customer Charge • By itself, it gives a meaningless price signal. Could a customer decide to not take service to avoid it? • To a rate analyst, it means establishing a closer link between cost allocation and rate design. However, a customer is not a rate analyst. • It is a charge that simply puts a utility in a bad light with its customers.